<p>I hear you, Mahras. According to Maboussin, the market aggregates all information from both short and long term players in such a way that equity prices reflect 12 month horizons. This in line with the high yearly turnover of most of the largest public mutual funds-they seek to maximize yoy returns as opposed to maximium return. I'm sure you've read the "equity risk premium puzzle."</p>
<p>Thus, a small player can, in my opinion-based on academic research, maximize their return by accepting greater volatility or beta on a yearly time line so as to maximize the multi-year return. Thus, low turnover is crucial, and looking for value where no short term catalyst exists for a fund manager to profit in the next 12 months will lead to realizing equities which have been afflicted by "hyperbolic discounting" in behavioral finance terms, i.e. the expectations of a few years out have been overly discounted because of a natural impulse to seek performance that will gratify fund holders in the next 12 months. The entire sell side research community is geared towards a 12 month price target paradigm.</p>
<p>On the other side, a shorter than 12 moth time line is useless for those not working with advanced algorithm trading and zero latency, priority trading. Wether or not stocks followed a random walk during Malkeil's time, they certaintly follow a random walk as far as the average retail investor is concerned. Some Arb potential does exist for the little guys where the big guys are too bloated to make any money, but often these trades are not even that profitable when taking into consideration margin costs, short term capital gains and transaction fees. </p>
<p>CMG and CMG-B is an arb that has been hyped for months in the blogosphere. When considering the cost of margin, this only proves my above point. By the time the spread closes, it doesn't matter that you "were right", in economic terms, your at a loss.</p>
<p>Also, I am a top ranked player on CAPS, and have been in the 99th percentile of all players everyday for over 6 months. Obviously, this is a few standard deviations out from what one would expect from randomness alone. However, this is partially because I'm not restricted to any market in my model porfolio, and can short at will, which is not always possible in the real world. Again, if my strategy faulters, a large loss wipes away all my hard work, as it's the geometric average that matters in the real world. Not enough traders look at low probability events.</p>
<p>Also, BIDU and RIMM are extremely overvalued. I'm gonna be pretty smug when BIDU inevitably repeats what the U.S. tech companies did in 2000.</p>